Exploratory Data Analysis

data <- read.csv("UCI-electricity/UCI_data.csv")

# Display the first few rows of the data
head(data)

##                  date       T1     RH_1       T2     RH_2       T3     RH_3
## 1 2016-04-19 20:30:00 22.20000 39.50000 20.56667 37.65667 22.23000 37.03000
## 2 2016-03-05 04:40:00 20.35667 37.12667 17.56667 40.23000 20.89000 37.66333
## 3 2016-03-14 12:40:00 20.92667 38.79000 21.10000 35.52667 21.60000 36.29000
## 4 2016-01-22 15:30:00 18.29000 38.90000 17.29000 39.26000 18.39000 39.32667
## 5 2016-02-10 00:40:00 22.29000 42.33333 21.60000 40.43333 22.66667 43.36333
## 6 2016-02-16 15:40:00 21.79000 34.73000 19.76000 35.40000 21.23000 35.20000
##         T4     RH_4       T5     RH_5       T6     RH_6    T7     RH_7    T8
## 1 22.31857 36.61000 20.63333 62.16667 8.823333 13.29667 21.00 29.31857 22.23
## 2 18.70000 36.26000 18.46333 43.56000 1.230000 78.96333 18.39 32.29000 21.00
## 3 21.00000 34.82667 18.10000 46.12667 6.033333 37.30000 21.29 27.56667 21.70
## 4 16.10000 38.79000 16.10000 47.70000 4.595000 94.94500 16.20 33.59000 17.78
## 5 19.10000 40.90000 19.29000 50.74500 2.530000 92.19000 20.00 40.00000 23.10
## 6 20.10000 34.76000 19.03333 38.40000 4.500000 46.70000 21.76 26.26333 22.60
##       RH_8       T9  RH_9    T_out Press_mm_hg   RH_out Windspeed Visibility
## 1 38.46333 19.92667 33.90 9.700000    766.1000 65.50000  3.500000   40.00000
## 2 43.40000 18.39000 41.09 0.300000    740.3333 99.00000  1.000000   41.33333
## 3 32.44750 19.39000 38.76 4.400000    768.4667 72.00000  6.000000   22.66667
## 4 41.66333 15.80000 39.20 3.350000    760.6000 82.00000  5.500000   41.00000
## 5 46.09000 18.89000 43.73 3.200000    738.9000 88.00000  7.333333   56.00000
## 6 32.66333 18.29000 40.20 4.633333    770.8667 52.33333  2.333333   40.00000
##    Tdewpoint       rv1       rv2 TARGET_energy
## 1  3.3500000 24.061869 24.061869            60
## 2  0.1000000  4.622052  4.622052            50
## 3 -0.2666667  5.635898  5.635898            80
## 4  0.5000000 49.216445 49.216445            40
## 5  1.4000000 47.617579 47.617579            60
## 6 -4.3333333 16.431218 16.431218            50

# Display the last few rows of the data
tail(data)

##                      date    T1  RH_1       T2     RH_2       T3     RH_3
## 19730 2016-02-09 18:40:00 22.39 41.40 21.96333 40.00000 22.39000 41.66333
## 19731 2016-01-27 06:30:00 20.10 44.90 19.39000 44.17250 19.70000 44.00000
## 19732 2016-04-14 08:50:00 21.39 41.09 19.23000 44.63333 22.79000 38.03000
## 19733 2016-04-25 08:30:00 20.89 36.50 18.00000 40.23000 22.13333 34.43333
## 19734 2016-03-04 06:50:00 20.79 36.70 18.79000 38.24500 21.00000 38.79000
## 19735 2016-02-16 10:20:00 19.79 39.20 18.23000 39.40000 21.55000 37.94000
##             T4     RH_4       T5     RH_5        T6     RH_6       T7     RH_7
## 19730 19.39000 40.70000 19.50000 49.18778  3.723333 92.30000 19.87889 40.90444
## 19731 19.06667 46.46667 17.61111 56.43333 10.263333 85.73000 18.20000 47.30222
## 19732 22.53333 39.36000 20.39000 44.59000  7.900000 62.29667 20.50000 34.55846
## 19733 20.70000 34.61200 20.00000 42.56000  6.433333 53.86667 21.29000 31.70000
## 19734 19.29000 36.11750 18.17143 51.50000  4.051429 79.75286 18.70000 32.48571
## 19735 19.23000 35.73000 18.20000 42.29000 -0.600000 86.06000 18.29000 29.89000
##             T8     RH_8      T9     RH_9     T_out Press_mm_hg   RH_out
## 19730 23.39000 44.72000 19.0000 43.09000  4.033333    735.7667  93.0000
## 19731 17.79000 49.59000 17.2300 49.59000 10.300000    757.3500  80.5000
## 19732 22.10000 40.79000 21.9725 38.34750  4.116667    754.3000 100.0000
## 19733 23.39000 37.79000 20.0000 36.32667  4.000000    754.8500  90.0000
## 19734 20.89000 44.82143 18.5000 41.72571  3.807143    740.5738  91.2381
## 19735 20.13333 37.96667 17.7900 37.43333  0.500000    772.2333  95.0000
##       Windspeed Visibility  Tdewpoint       rv1       rv2 TARGET_energy
## 19730  4.000000       40.0  2.9666667 19.390282 19.390282           140
## 19731 10.000000       34.5  7.0000000  2.233453  2.233453            40
## 19732  1.000000       33.0  4.1166667 18.511994 18.511994           100
## 19733  3.000000       25.5  2.4500000  6.176545  6.176545            70
## 19734  7.285714       58.5  2.4547619  4.780297  4.780297            50
## 19735  1.333333       62.0 -0.2666667 23.204486 23.204486           280

# Display the structure of the data
str(data)

## 'data.frame':    19735 obs. of  28 variables:
##  $ date         : chr  "2016-04-19 20:30:00" "2016-03-05 04:40:00" "2016-03-14 12:40:00" "2016-01-22 15:30:00" ...
##  $ T1           : num  22.2 20.4 20.9 18.3 22.3 ...
##  $ RH_1         : num  39.5 37.1 38.8 38.9 42.3 ...
##  $ T2           : num  20.6 17.6 21.1 17.3 21.6 ...
##  $ RH_2         : num  37.7 40.2 35.5 39.3 40.4 ...
##  $ T3           : num  22.2 20.9 21.6 18.4 22.7 ...
##  $ RH_3         : num  37 37.7 36.3 39.3 43.4 ...
##  $ T4           : num  22.3 18.7 21 16.1 19.1 ...
##  $ RH_4         : num  36.6 36.3 34.8 38.8 40.9 ...
##  $ T5           : num  20.6 18.5 18.1 16.1 19.3 ...
##  $ RH_5         : num  62.2 43.6 46.1 47.7 50.7 ...
##  $ T6           : num  8.82 1.23 6.03 4.59 2.53 ...
##  $ RH_6         : num  13.3 79 37.3 94.9 92.2 ...
##  $ T7           : num  21 18.4 21.3 16.2 20 ...
##  $ RH_7         : num  29.3 32.3 27.6 33.6 40 ...
##  $ T8           : num  22.2 21 21.7 17.8 23.1 ...
##  $ RH_8         : num  38.5 43.4 32.4 41.7 46.1 ...
##  $ T9           : num  19.9 18.4 19.4 15.8 18.9 ...
##  $ RH_9         : num  33.9 41.1 38.8 39.2 43.7 ...
##  $ T_out        : num  9.7 0.3 4.4 3.35 3.2 ...
##  $ Press_mm_hg  : num  766 740 768 761 739 ...
##  $ RH_out       : num  65.5 99 72 82 88 ...
##  $ Windspeed    : num  3.5 1 6 5.5 7.33 ...
##  $ Visibility   : num  40 41.3 22.7 41 56 ...
##  $ Tdewpoint    : num  3.35 0.1 -0.267 0.5 1.4 ...
##  $ rv1          : num  24.06 4.62 5.64 49.22 47.62 ...
##  $ rv2          : num  24.06 4.62 5.64 49.22 47.62 ...
##  $ TARGET_energy: int  60 50 80 40 60 50 60 400 80 100 ...

library(striprtf)

# Specify the path to your RTF file
file_path <- "UCI-electricity/description.rtf"

# Read the RTF file
rtf_content <- read_rtf(file_path)

# Display the content
cat(rtf_content)

## *| Attribute Information:date time year-month-day hour:minute:second T1, Temperature in kitchen area, in Celsius RH_1, Humidity in kitchen area, in % T2, Temperature in living room area, in Celsius RH_2, Humidity in living room area, in % T3, Temperature in laundry room area RH_3, Humidity in laundry room area, in % T4, Temperature in office room, in Celsius RH_4, Humidity in office room, in % T5, Temperature in bathroom, in Celsius RH_5, Humidity in bathroom, in % T6, Temperature outside the building (north side), in Celsius RH_6, Humidity outside the building (north side), in % T7, Temperature in ironing room , in Celsius RH_7, Humidity in ironing room, in % T8, Temperature in teenager room 2, in Celsius RH_8, Humidity in teenager room 2, in % T9, Temperature in parents room, in Celsius RH_9, Humidity in parents room, in % To, Temperature outside (from Chievres weather station), in Celsius Pressure (from Chievres weather station), in mm Hg RH_out, Humidity outside (from Chievres weather station), in % Wind speed (from Chievres weather station), in m/s Visibility (from Chievres weather station), in km Tdewpoint (from Chievres weather station), 슰C rv1, Random variable 1, nondimensional rv2, Random variable 2, nondimensional TARGET_Energy, energy use of Appliances and light fixtures in the house in Wh  |

# Count of missing values in each column
colSums(is.na(data))

##          date            T1          RH_1            T2          RH_2 
##             0             0             0             0             0 
##            T3          RH_3            T4          RH_4            T5 
##             0             0             0             0             0 
##          RH_5            T6          RH_6            T7          RH_7 
##             0             0             0             0             0 
##            T8          RH_8            T9          RH_9         T_out 
##             0             0             0             0             0 
##   Press_mm_hg        RH_out     Windspeed    Visibility     Tdewpoint 
##             0             0             0             0             0 
##           rv1           rv2 TARGET_energy 
##             0             0             0

# Percentage of missing values in each column
colMeans(is.na(data)) * 100

##          date            T1          RH_1            T2          RH_2 
##             0             0             0             0             0 
##            T3          RH_3            T4          RH_4            T5 
##             0             0             0             0             0 
##          RH_5            T6          RH_6            T7          RH_7 
##             0             0             0             0             0 
##            T8          RH_8            T9          RH_9         T_out 
##             0             0             0             0             0 
##   Press_mm_hg        RH_out     Windspeed    Visibility     Tdewpoint 
##             0             0             0             0             0 
##           rv1           rv2 TARGET_energy 
##             0             0             0

# Set up the plotting area for multiple plots
par(mfrow=c(3, 1), mar=c(4, 4, 2, 1), oma=c(0, 0, 2, 0))

# Plot the histograms for temperature variables
par(mfrow=c(2, 5))
hist(data$T1, main="Kitchen", xlab="T1 (Celsius)", col="blue")
hist(data$T2, main="Living Room", xlab="T2 (Celsius)", col="green")
hist(data$T3, main="Laundry Room", xlab="T3 (Celsius)", col="red")
hist(data$T4, main="Office Room", xlab="T4 (Celsius)", col="purple")
hist(data$T5, main="Bathroom", xlab="T5 (Celsius)", col="orange")
hist(data$T6, main="Outside North", xlab="T6 (Celsius)", col="yellow")
hist(data$T7, main="Ironing Room", xlab="T7 (Celsius)", col="pink")
hist(data$T8, main="Teenager Room 2", xlab="T8 (Celsius)", col="cyan")
hist(data$T9, main="Parents Room", xlab="T9 (Celsius)", col="brown")
hist(data$T_out, main="Outside Chievres", xlab="T_out (Celsius)", col="grey")

# Add an overall title for the temperature histograms
mtext("Histograms of Temperature Variables", outer = TRUE, line = -2, cex = 1.5)

# Plot the histograms for humidity variables
par(mfrow=c(2, 5))
hist(data$RH_1, main="Kitchen", xlab="RH_1 (%)", col="blue")
hist(data$RH_2, main="Living Room", xlab="RH_2 (%)", col="green")
hist(data$RH_3, main="Laundry Room", xlab="RH_3 (%)", col="red")
hist(data$RH_4, main="Office Room", xlab="RH_4 (%)", col="purple")
hist(data$RH_5, main="Bathroom", xlab="RH_5 (%)", col="orange")
hist(data$RH_6, main="Outside North", xlab="RH_6 (%)", col="yellow")
hist(data$RH_7, main="Ironing Room", xlab="RH_7 (%)", col="pink")
hist(data$RH_8, main="Teenager Room 2", xlab="RH_8 (%)", col="cyan")
hist(data$RH_9, main="Parents Room", xlab="RH_9 (%)", col="brown")
hist(data$RH_out, main="Outside Chievres", xlab="RH_out (%)", col="grey")

# Add an overall title for the humidity histograms
mtext("Histograms of Humidity Variables", outer = TRUE, line = -22, cex = 1.5)

# Plot the histograms for additional numerical variables
par(mfrow=c(2, 3))
hist(data$Press_mm_hg, main="Pressure (Chievres)", xlab="Press_mm_hg (mm Hg)", col="blue")
hist(data$Windspeed, main="Wind Speed (Chievres)", xlab="Windspeed (m/s)", col="green")
hist(data$Visibility, main="Visibility (Chievres)", xlab="Visibility (km)", col="red")
hist(data$Tdewpoint, main="Dewpoint (Chievres)", xlab="Tdewpoint (Celsius)", col="purple")
hist(data$rv1, main="Random Variable 1", xlab="rv1", col="orange")
hist(data$rv2, main="Random Variable 2", xlab="rv2", col="yellow")

hist(data$TARGET_energy, main="Energy Use (Wh)", xlab="TARGET_energy (Wh)", col="pink")

# Add an overall title for the additional numerical variable histograms
mtext("Histograms of Additional Numerical Variables", outer = TRUE, line = -42, cex = 1.5)

Analysis of Temperature Variables

• Kitchen (T1):
  • Distribution is approximately normal with a peak around 22°C.
  • Most values range between 18°C and 26°C, indicating a stable temperature range in the kitchen.
• Living Room (T2):
  • Slightly right-skewed distribution with a peak around 21°C.
  • Values range from 16°C to 28°C, showing some variability but generally centered around comfortable room temperature.
• Laundry Room (T3):
  • Approximately normal distribution with a peak around 24°C.
  • Values range between 18°C and 30°C, suggesting consistent conditions.
• Office Room (T4):
  • Normal distribution with a peak around 21°C.
  • Values range between 16°C and 24°C, indicating good temperature control.
• Bathroom (T5):
  • Approximately normal distribution with a peak around 22°C.
  • Values range between 16°C and 24°C, suggesting stable temperature control.
• Outside North (T6):
  • Distribution is right-skewed with a peak around 10°C.
  • Values range from 0°C to 30°C, reflecting wider external temperature variations.
• Ironing Room (T7):
  • Approximately normal distribution with a peak around 21°C.
  • Values range between 16°C and 24°C, indicating consistent temperature.
• Teenager Room 2 (T8):
  • Normal distribution with a peak around 22°C.
  • Values range between 16°C and 24°C, showing consistent conditions.
• Parents Room (T9):
  • Approximately normal distribution with a peak around 21°C.
  • Values range from 16°C to 24°C, suggesting stable conditions.
• Outside (Chievres, T_out):
  • Right-skewed distribution with a peak around 10°C.
  • Values range from -5°C to 25°C, showing greater external temperature variation.

Analysis of Humidity Variables

• Kitchen (RH_1):
  • Normal distribution with a peak around 40%.
  • Values range between 30% and 50%, indicating stable humidity.
• Living Room (RH_2):
  • Normal distribution with a peak around 35%.
  • Values range between 20% and 40%, showing stable conditions.
• Laundry Room (RH_3):
  • Normal distribution with a peak around 40%.
  • Values range between 30% and 50%, suggesting consistent conditions.
• Office Room (RH_4):
  • Normal distribution with a peak around 35%.
  • Values range between 25% and 45%, indicating stable humidity.
• Bathroom (RH_5):
  • Right-skewed distribution with a peak around 60%.
  • Values range between 40% and 80%, reflecting variability.
• Outside North (RH_6):
  • Bi-modal distribution with peaks around 15% and 80%.
  • Values range from 0% to 100%, reflecting external humidity variability.
• Ironing Room (RH_7):
  • Normal distribution with a peak around 35%.
  • Values range between 25% and 45%, indicating consistent conditions.
• Teenager Room 2 (RH_8):
  • Normal distribution with a peak around 45%.
  • Values range between 30% and 60%, showing stable humidity.
• Parents Room (RH_9):
  • Normal distribution with a peak around 45%.
  • Values range from 30% to 50%, suggesting stable conditions.
• Outside (Chievres, RH_out):
  • Right-skewed distribution with a peak around 80%.
  • Values range from 20% to 100%, reflecting greater external humidity variation.

Analysis of Additional Numerical Variables

• Pressure (Chievres):
  • Approximately normal distribution with a peak around 760 mm Hg.
  • Values range between 730 mm Hg and 770 mm Hg, indicating stable atmospheric pressure.
• Wind Speed (Chievres):
  • Right-skewed distribution with a peak around 2 m/s.
  • Values range from 0 to 14 m/s, reflecting variability in wind speed.
• Visibility (Chievres):
  • Bi-modal distribution with peaks around 30 km and 40 km.
  • Values range from 0 to 70 km, showing variability in visibility.
• Dewpoint (Chievres):
  • Approximately normal distribution with a peak around 5°C.
  • Values range between -5°C and 15°C, indicating variability in dew point.
• Random Variable 1 (rv1):
  • Uniform distribution across the range.
  • Values range between 0 and 50, suggesting a random distribution.
• Random Variable 2 (rv2):
  • Uniform distribution across the range.
  • Values range between 0 and 50, suggesting a random distribution.
• Energy Use (Wh):
  • Right-skewed distribution with a peak around 50 Wh.
  • Values range from 0 to 800 Wh, showing variability in energy usage.

# Compute the correlation matrix
cor_matrix <- cor(data[, sapply(data, is.numeric)], use="complete.obs")

# Install and load the corrplot package
#install.packages("corrplot")
library(corrplot)

## corrplot 0.92 loaded

# Visualize the correlation matrix
corrplot(cor_matrix, method="circle", tl.cex = 0.5, tl.col = "black", tl.srt = 45)

## Correlation Analysis

The correlation matrix plot above visualizes the relationships between the numerical variables in the dataset. Here are the key observations:

• Strong Positive Correlations:
  • T1, T2, T3, T4, T5, T7, T8, T9: These temperature variables are strongly positively correlated with each other (correlation coefficients close to 1). This indicates that temperatures across different rooms in the house tend to increase and decrease together.
  • RH_1, RH_2, RH_3, RH_4, RH_5, RH_7, RH_8, RH_9: These humidity variables are also strongly positively correlated with each other, suggesting that humidity levels across different rooms tend to move in tandem.
• Negative Correlations:
  • T1 and RH_1, T2 and RH_2, etc.: There are moderate to strong negative correlations between temperature and humidity in the same room, indicating that as temperature increases, humidity decreases, and vice versa.
• External Temperature (T_out):
  • T_out and T1, T2, etc.: External temperature (T_out) has moderate positive correlations with internal temperatures, suggesting that changes in external temperature influence internal temperatures.
• Pressure (Press_mm_hg):
  • Pressure and Humidity: There is a slight negative correlation between pressure and humidity levels (RH_out), indicating that higher pressure may be associated with lower humidity levels.
• Wind Speed (Windspeed):
  • Windspeed and Temperature: Wind speed shows slight negative correlations with internal temperatures, suggesting that higher wind speeds might slightly reduce indoor temperatures.
• Visibility:
  • Visibility and External Variables: Visibility shows slight positive correlations with external temperature and slight negative correlations with external humidity.
• Energy Use (TARGET_energy):
  • TARGET_energy and Internal Variables: Energy use shows weak to moderate positive correlations with internal temperatures (e.g., T1, T2) and negative correlations with internal humidity levels (e.g., RH_1, RH_2). This indicates that higher energy use is associated with higher temperatures and lower humidity levels.

These observations provide insights into how different environmental variables interact within the house and with external conditions. The correlations between temperatures, humidity, and energy use can help in understanding the energy dynamics and potential factors influencing energy consumption in the house.

# Install necessary packages if not already installed
#install.packages("GGally")
#install.packages("dplyr")

# Load the packages
library(GGally)

## Loading required package: ggplot2

## Registered S3 method overwritten by 'GGally':
##   method from   
##   +.gg   ggplot2

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

# Select the columns for temperature, humidity, and additional numerical variables
temperature_vars <- data %>% select(T1, T2, T3, T4, T5, T6, T7, T8, T9, T_out)
humidity_vars <- data %>% select(RH_1, RH_2, RH_3, RH_4, RH_5, RH_6, RH_7, RH_8, RH_9, RH_out)
additional_vars <- data %>% select(Press_mm_hg, Windspeed, Visibility, Tdewpoint, rv1, rv2, TARGET_energy)

# Create pair plot for temperature variables
ggpairs(temperature_vars, title = "Pair Plot for Temperature Variables")

# Create pair plot for humidity variables
ggpairs(humidity_vars, title = "Pair Plot for Humidity Variables")

# Create pair plot for additional numerical variables
ggpairs(additional_vars, title = "Pair Plot for Additional Numerical Variables")

# Box plots for temperature variables
par(mfrow=c(2, 5)) # Arrange plots in a 2x5 grid
boxplot(data$T1, main="Kitchen", col="blue")
boxplot(data$T2, main="Living Room", col="green")
boxplot(data$T3, main="Laundry Room", col="red")
boxplot(data$T4, main="Office Room", col="purple")
boxplot(data$T5, main="Bathroom", col="orange")
boxplot(data$T6, main="Outside North", col="yellow")
boxplot(data$T7, main="Ironing Room", col="pink")
boxplot(data$T8, main="Teenager Room 2", col="cyan")
boxplot(data$T9, main="Parents Room", col="brown")
boxplot(data$T_out, main="Outside (Chievres)", col="grey")

# Box plots for humidity variables
par(mfrow=c(2, 5)) # Arrange plots in a 2x5 grid
boxplot(data$RH_1, main="Kitchen", col="blue")
boxplot(data$RH_2, main="Living Room", col="green")
boxplot(data$RH_3, main="Laundry Room", col="red")
boxplot(data$RH_4, main="Office Room", col="purple")
boxplot(data$RH_5, main="Bathroom", col="orange")
boxplot(data$RH_6, main="Outside North", col="yellow")
boxplot(data$RH_7, main="Ironing Room", col="pink")
boxplot(data$RH_8, main="Teenager Room 2", col="cyan")
boxplot(data$RH_9, main="Parents Room", col="brown")
boxplot(data$RH_out, main="Outside (Chievres)", col="grey")

#install.packages("ggplot2")
#install.packages("dplyr")
#install.packages("tidyr")

# Load the packages
library(ggplot2)
library(dplyr)
library(tidyr)

# Convert the 'date' column to datetime format
data$date <- as.POSIXct(data$date, format="%Y-%m-%d %H:%M:%S")

# Create long format for temperature variables
temperature_vars <- data %>%
  select(date, T1, T2, T3, T4, T5, T6, T7, T8, T9, T_out) %>%
  pivot_longer(cols = -date, names_to = "variable", values_to = "value")

# Plot time series for temperature variables
ggplot(temperature_vars, aes(x=date, y=value, color=variable)) +
  geom_line() +
  labs(title="Temperature Time Series", y="Temperature (Celsius)", x="Date") +
  theme_minimal() +
  theme(legend.position="bottom") +
  scale_color_manual(values=c("blue", "green", "red", "purple", "orange", "yellow", "pink", "cyan", "brown", "grey"))

# Create long format for humidity variables
humidity_vars <- data %>%
  select(date, RH_1, RH_2, RH_3, RH_4, RH_5, RH_6, RH_7, RH_8, RH_9, RH_out) %>%
  pivot_longer(cols = -date, names_to = "variable", values_to = "value")

# Plot time series for humidity variables
ggplot(humidity_vars, aes(x=date, y=value, color=variable)) +
  geom_line() +
  labs(title="Humidity Time Series", y="Humidity (%)", x="Date") +
  theme_minimal() +
  theme(legend.position="bottom") +
  scale_color_manual(values=c("blue", "green", "red", "purple", "orange", "yellow", "pink", "cyan", "brown", "grey"))

# Plot time series for energy use
ggplot(data, aes(x=date, y=TARGET_energy)) +
  geom_line(color="blue") +
  labs(title="Energy Use Time Series", y="Energy Use (Wh)", x="Date") +
  theme_minimal()

### 1 Humidity Time Series

Critical Analysis

The humidity time series plot above visualizes the variations in humidity levels across different rooms and the outside environment over time. Here are the key observations:

• High Variability:
  • RH_out: There are significant fluctuations in the outside humidity levels, with frequent spikes and dips, indicating variable weather conditions.
  • RH_1 to RH_9: Humidity levels inside the house are relatively stable, suggesting effective climate control.
• Patterns:
  • RH_5 and RH_7: Certain rooms show different humidity patterns, which could be due to localized factors like the presence of water sources or ventilation differences.
  • Indoor vs. Outdoor: The stability of indoor humidity compared to outdoor suggests good insulation and effective use of HVAC systems.

2. Temperature Time Series

Critical Analysis

The temperature time series plot above visualizes the variations in temperature levels across different rooms and the outside environment over time. Here are the key observations:

• Seasonal Patterns:
  • T_out: External temperature shows clear seasonal patterns, with temperatures fluctuating more significantly during certain periods.
  • T1 to T9: Indoor temperatures remain relatively stable, indicating effective temperature control within the house.
• Room Variability:
  • T3, T5: Some rooms show more variability in temperature, potentially due to their location within the house or varying usage patterns.
• Insulation Effect:
  • The difference between indoor and outdoor temperature patterns suggests good insulation, which helps in maintaining a stable indoor environment despite external temperature changes.

3. Energy Use Time Series

Critical Analysis

The energy use time series plot above visualizes the variations in energy consumption over time. Here are the key observations:

• High Variability:
  • Energy use shows significant fluctuations, with frequent spikes that likely correspond to peak usage times or specific high-energy activities.
• Cyclical Patterns:
  • There appears to be a cyclical pattern in energy use, potentially corresponding to daily or weekly usage cycles, indicating regular consumption behaviors.
• Peak Usage:
  • The spikes in energy use highlight peak consumption periods, which could be targeted for energy-saving measures to improve overall efficiency.

Conclusion

The time series plots provide valuable insights into the behavior of humidity, temperature, and energy use in the building. Understanding these patterns is crucial for developing predictive models and implementing energy-saving measures.

Further analysis could include:

• Correlating temperature and humidity with energy use to identify key drivers.
• Exploring anomalies and outliers to improve the reliability of the predictive models.
• Implementing feature engineering techniques to capture seasonal and cyclical patterns effectively.

By leveraging these insights, we can develop robust models to predict energy use and optimize energy consumption in low-energy buildings.

#install.packages("forecast")
library(forecast)

## Registered S3 method overwritten by 'quantmod':
##   method            from
##   as.zoo.data.frame zoo

# Decompose energy use time series
energy_ts <- ts(data$TARGET_energy, frequency=24) # Assuming hourly data
decomposed <- stl(energy_ts, s.window="periodic")
plot(decomposed)

### Critical Analysis

The time series decomposition plot above visualizes the components of the energy use data over time. Here are the key observations:

• Data:
  • The raw energy use data shows significant fluctuations with frequent spikes, indicating variability in energy consumption.
• Seasonal:
  • The seasonal component captures the repeating patterns in the data. The plot shows a strong and regular seasonal pattern, suggesting that there are consistent daily or weekly cycles in energy use.
• Trend:
  • The trend component shows the long-term direction of the energy use. There appears to be an increasing trend over time, indicating that overall energy consumption is rising.
• Remainder:
  • The remainder component captures the residuals after removing the seasonal and trend components. This plot shows the noise and random fluctuations in the data. The high variability in the remainder indicates that there are other unexplained factors influencing energy use.

Conclusion

The decomposition plot provides valuable insights into the underlying patterns in the energy use data: - The strong seasonal component suggests regular cycles, which could be leveraged for predictive modeling. - The increasing trend indicates a long-term rise in energy consumption, which might require addressing underlying causes. - The high variability in the remainder component suggests the presence of additional factors affecting energy use, warranting further investigation.

# Lagged correlation
library(dplyr)
lagged_data <- data %>%
  mutate(TARGET_energy_lag1 = lag(TARGET_energy, 1),
         TARGET_energy_lag24 = lag(TARGET_energy, 24)) # Example lags

cor(lagged_data[, c("TARGET_energy", "TARGET_energy_lag1", "TARGET_energy_lag24")], use="complete.obs")

##                     TARGET_energy TARGET_energy_lag1 TARGET_energy_lag24
## TARGET_energy         1.000000000        0.005404513        -0.012206256
## TARGET_energy_lag1    0.005404513        1.000000000         0.007792859
## TARGET_energy_lag24  -0.012206256        0.007792859         1.000000000

library(ggplot2)
#install.packages("reshape2")
library(reshape2)

## 
## Attaching package: 'reshape2'

## The following object is masked from 'package:tidyr':
## 
##     smiths

# Heatmap of correlation matrix
cor_matrix <- cor(data[, sapply(data, is.numeric)], use="complete.obs")
melted_cor_matrix <- melt(cor_matrix)
ggplot(data = melted_cor_matrix, aes(x=Var1, y=Var2, fill=value)) +
  geom_tile() +
  scale_fill_gradient2(low="blue", high="red", mid="white", midpoint=0) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
  labs(title="Correlation Heatmap")

## Critical Analysis

Correlation Heatmap

The correlation heatmap above visualizes the relationships between various numerical variables in the dataset. Here are the key observations:

• Strong Positive Correlations:
  • Temperature Variables (T1 to T9, T_out): These temperature variables are strongly positively correlated with each other, indicating that temperatures across different rooms in the house tend to increase and decrease together.
  • Humidity Variables (RH_1 to RH_9): These humidity variables also show strong positive correlations with each other, suggesting that humidity levels across different rooms tend to move in tandem.
• Negative Correlations:
  • Temperature and Humidity: There are moderate to strong negative correlations between temperature and humidity in the same room (e.g., T1 and RH_1, T2 and RH_2), indicating that as temperature increases, humidity decreases, and vice versa.
• External Temperature (T_out):
  • T_out and Internal Temperatures: External temperature (T_out) has moderate positive correlations with internal temperatures, suggesting that changes in external temperature influence internal temperatures.
• Pressure (Press_mm_hg):
  • Pressure and Humidity: There is a slight negative correlation between pressure and humidity levels (RH_out), indicating that higher pressure may be associated with lower humidity levels.
• Wind Speed (Windspeed):
  • Windspeed and Temperature: Wind speed shows slight negative correlations with internal temperatures, suggesting that higher wind speeds might slightly reduce indoor temperatures.
• Visibility:
  • Visibility and External Variables: Visibility shows slight positive correlations with external temperature and slight negative correlations with external humidity.
• Energy Use (TARGET_energy):
  • Weak Correlations: Energy use (TARGET_energy) shows weak correlations with most other variables, indicating that other factors might be influencing energy consumption more strongly.

Lagged Correlation Analysis

The table below shows the correlation between energy use and its lagged values:

	TARGET_energy	TARGET_energy_lag1	TARGET_energy_lag24
TARGET_energy	1.000000000	0.005404513	-0.012206256
TARGET_energy_lag1	0.005404513	1.000000000	0.007792859
TARGET_energy_lag24	-0.012206256	0.007792859	1.000000000

Observations: - Weak Correlations with Lagged Values: - The correlation between current energy use and energy use lagged by 1 hour (TARGET_energy_lag1) is very weak (0.005). - The correlation between current energy use and energy use lagged by 24 hours (TARGET_energy_lag24) is also very weak (-0.012).

Conclusion

The correlation heatmap and lagged correlation analysis provide valuable insights into the relationships between different variables in the dataset: - The strong correlations among temperature variables and among humidity variables indicate consistent environmental conditions within the house. - The weak correlations between energy use and its lagged values suggest that current energy consumption is not strongly dependent on past values, indicating the influence of other factors. - The negative correlations between temperature and humidity highlight the interplay between these two environmental factors.